This syllabus is a working draft. Logistics, readings, assignments, and the weekly schedule will all change before the semester starts.
This is the home page of the course syllabus, which outlines all the instructions, logistics, and expectations for the course. The syllabus also has a schedule page, which details the readings and work for each week.
What makes someone an expert? And why does anyone believe them?
Every scholar depends on knowledge they can’t check for themselves: a colleague’s archival work, a translation, an edition, a dataset, a footnote. Expertise is how we manage that dependence, and trust is what makes it work. Neither is natural or permanent. Both have been built over time by people, processes, and institutions, often in response to a new technology that disrupted old arrangements.
Generative AI is the latest of these technologies, and it is routinely called unprecedented. This seminar argues that it isn’t, and that historians are especially well equipped to say why. Plato worried that writing would give people the appearance of wisdom without the substance. Early print multiplied errors as fast as it multiplied books, and it took generations of correctors, privileges, and registers before readers trusted a printed page. Experimental science had to recruit credible witnesses before anyone believed a demonstration. Statistics, mechanical calculation, and keyword search each promised to replace expert judgment, and each ended up relocating it. In every case, trust did not arrive with the technology. People had to manufacture it.
The course moves in three parts. First we build a working vocabulary for expertise and trust from philosophy, sociology, and the history of science. Then we work through historical disruptions of expertise, from medieval charters to digitized newspapers. Finally we turn to AI itself: what language models actually do, how their interfaces perform authority, and what happens to scholarly expertise when anyone can produce in seconds something that looks like a historian’s work.
If a machine can produce something that looks like expertise, what was expertise in the first place?
Four questions run through every week of the course, whether the expert in question is a medieval scribe, a Victorian statistician, or a chatbot.
This seminar is designed for history graduate students, and students from every discipline are welcome. You don’t need any technical background, and you don’t need to like technology. We build the few technical ideas we need from the ground up, and only as far as they help with historical and epistemological questions.
Skepticism about AI is an asset here, as long as you can explain it. What you do need is a field you care about: the best test of anything we read is what it says about expertise you already have.
There are no required books to purchase. We use a tool called Zotero to manage PDFs of the readings.
To get connected, follow the Zotero set up instructions, noting the links below.
Group library links will be posted here once the course library is set up.
These are draft assignments and weights.
Come having read, and come ready to talk. Each of you will also co-lead discussion once or twice during the semester: frame the week’s readings, put them in conversation with earlier weeks, and bring the questions you think are worth our time.
A short response (~500 words) posted before each seminar. Don’t summarize. Put the week’s readings in conversation with each other or with earlier weeks, and end with a few questions you want to discuss.
Choose a live debate in a field you know well. Ask an AI system the kinds of questions you’d expect on a comprehensive exam, and check everything it tells you, especially the citations. Then try the same thing in a field you don’t know. Write up (~2,500 words) where the system was reliable, where it failed, what its failures reveal about the field’s literature, and how you could tell. The comparison between your two fields is the point: what could you catch, and why?
Choose a historical technology of trust from the course (the witness list, the colophon, the footnote, the referee report, the credential, the published table) and a feature of an AI interface that does similar work (cited sources, confidence language, visible “reasoning,” personas, disclaimers). Analyze how each asks to be trusted and what it offers in return (~2,000 words). Optionally, add a short design brief for an interface that earns trust rather than performing it.
Choose one:
Everyone presents their project to the seminar at the end of the semester.
All assignments are graded as A, B, C, D, F. Canvas calculates these into percentages in the Gradebook, so you can always see your current percent and letter grade based on work you’ve submitted.
| Percent | Grade |
|---|---|
| 98+ | A+ |
| 93-97 | A |
| 90-92 | A- |
| 87-89 | B+ |
| 83-86 | B |
| 80-82 | B- |
| 77-79 | C+ |
| 73-76 | C |
If life gets overwhelming during the course and you question whether you can continue, please reach out. We can discuss formal or informal accommodations, deadline adjustments, or other support needed for you to complete the course successfully enough for academic progress and to keep funding.
You should be familiar with UNM’s Policy on Academic Dishonesty and the Student Code of Conduct, which outline academic misconduct defined as plagiarism, cheating, fabrication, or facilitating any such act.
UNM is committed to providing equitable access to learning opportunities for students with documented disabilities. As your instructor, it is my objective to facilitate an inclusive classroom setting, in which students have full access and opportunity to participate. To engage in a confidential conversation about the process for requesting reasonable accommodations for this class and/or program, please contact Accessibility Resource Center at arcsrvs@unm.edu or 505-277-3506.
Our classroom and university should foster mutual respect, kindness, and support. If you have concerns about discrimination, harassment, or violence, please seek support and report incidents. Find confidential services at LoboRESPECT Advocacy Center, the Women’s Resource Center, and the Arcoiris Center. UNM prohibits discrimination on the basis of sex (including gender, sex stereotyping, gender expression, and gender identity). All instructors are “responsible employees” who must communicate reports of sexual harassment, sexual misconduct and sexual violence to Compliance, Ethics and Equal Opportunity. For more information, please see UAP 2720 and UAP 2740.
These indicate something you have to DO or TURN IN.
These indicate something you should be aware of—usually an upcoming assignment—but isn’t anything you need to immediately do.
These indicate something that is important to know, but isn’t time sensitive.